Guth Labs News · Discover
News
Current public News snapshot, its cited links and source refreshes, and published guide updates. This index does not fact-check entries or cover the full Guth catalog. New shows dated news and guide entries from the past seven days; routine source refreshes are excluded. All is the default.
426 resources in this view. Showing 101–200. This live snapshot can change between pages.
Publisher update
Plan or Retrieve First? A Scale Study of Tool Discovery Architectures
Read at Apple Machine Learning Research.
Publisher update
Selection-Based Structured Reasoning: Toward Efficient Multimodal Search Agents
Read at Apple Machine Learning Research.
Publisher update
Sullivan & Cromwell brings legal expertise into agreement review
Read at OpenAI News.
Publisher update
Welcome Muse Spark Open: open-weight frontier multimodal mixture-of-experts model from Meta
Read at Hugging Face Blog.
Guth briefing · AI-written
Raycast 2.8 for Mac targets speed in file search, extension installs and startup
Raycast has published version 2.8 of its Mac launcher, and the company describes the release as being about speed.
Guth briefing · AI-written
Google pledges $3 million in compute credits for European researchers
Up to 200 scientists would receive access to tools including AlphaFold 3, but the announcement leaves selection and timing unspecified.
Publisher update
Build self-improving document processing using Amazon Bedrock AgentCore
Read at AWS Machine Learning Blog.
Publisher update
Putting Earth AI to work: Geospatial agents for our planet's biggest challenges
Read at Google Research Blog.
Publisher update
Research@ 2026: 10 ways we're amplifying human ingenuity
Read at Google Research Blog.
Publisher update
HiRO-ACEv2: Simulating extreme weather at 3 km resolution for a fraction of the cost
Read at Hugging Face Blog.
Publisher update
Optimizing Agent Reliability and Tool Calling Efficiency for Chip Design Using NVIDIA NeMo Gym
Read at NVIDIA Developer Blog.
Publisher update
Generate presentations with Amazon Bedrock AgentCore
Read at AWS Machine Learning Blog.
Publisher update
How Tamnoon built AI-powered cloud security remediation with Amazon Bedrock
Read at AWS Machine Learning Blog.
Publisher update
Deploy and evaluate models with Amazon SageMaker JumpStart
Read at AWS Machine Learning Blog.
Publisher update
How Block is building AI-powered security with OpenAI technology
Read at OpenAI News.
Publisher update
European Cities Deploy NVIDIA-Powered AI Agents to Run Smarter Urban Operations
Read at NVIDIA Blog.
Publisher update
Division of Program Coordination, Planning, and Strategic Initiatives, Office of the Director Notice of Proposed Reorganization
The Division of Program Coordination, Planning, and Strategic Initiatives (DPCPSI) in the Office of the Director (OD), National Institutes of Health (NIH) is seeking public comment regarding its proposal to establish the Office of the NIH Chief Artificial Intelligence Officer…
Publisher update
Google DeepMind and Singapore Government collaborate to build fine-tuned AI model for scam detection
Read at Google DeepMind Blog.
Guth briefing · AI-written
GitHub adds a fresh-policy check for Copilot startup
A separate sandbox version gate requires users on older clients to update before they can resume using Copilot.
Publisher update
Code-Switching Detection via Language-Selective Dimensions in Speech Foundation Models
Read at Apple Machine Learning Research.
Publisher update
The Design Space of Tri-Modal Masked Diffusion Models
Read at Apple Machine Learning Research.
Publisher update
How to Rewrite Apache Spark UDFs for GPU Acceleration with AI Agents
Read at NVIDIA Developer Blog.
Publisher update
MIT Lincoln Laboratory hosts National Security Commission on Emerging Biotechnology
Read at MIT News.
Publisher update
Building agents that transact with Anchor Browser and AgentCore payments
Read at AWS Machine Learning Blog.
Publisher update
Manage private pricing offers for Amazon Bedrock with managed entitlements
Read at AWS Machine Learning Blog.
Publisher update
Natural language security queries at Orca Security using Amazon Bedrock
Read at AWS Machine Learning Blog.
Publisher update
How ShipStation Global built an AI IT help desk on Amazon Bedrock AgentCore
Read at AWS Machine Learning Blog.
Publisher update
Build long-horizon agents with Grok 4.7 on Amazon Bedrock
Read at AWS Machine Learning Blog.
Publisher update
Mixture-of-Experts kernel engineering on Amazon SageMaker AI
Read at AWS Machine Learning Blog.
Guth briefing · AI-written
GitHub adds explicit signals for withdrawn Dependabot alerts
Existing webhook subscribers will receive withdrawal events automatically, with the changes also planned for GitHub Enterprise Server 3.24.
Publisher update
NVIDIA Run:ai Is Now Open Source as Part of NVIDIA DSX AI Factory Platform
Read at NVIDIA Developer Blog.
Publisher update
AISN #83: Six AI Companies Sign White House Accord
Read at Center for AI Safety Newsletter.
Publisher update
All Systems Go: Global PC App Ecosystem Ready for NVIDIA RTX Spark
Read at NVIDIA Blog.
Publisher update
How Molmo 2’s open release helped an MIT researcher rethink fish tracking
Read at Allen Institute for AI.
Guth briefing · AI-written
CircleCI offers up to $20,000 in free compute credits for qualifying startups
CircleCI’s new startup program gives qualifying early‑stage companies up to $20,000 in free compute credits, 6,000 build minutes and 30x concurrency.
Guth briefing · AI-written
AI models break spatial rules in GeoSpatialBench tests, authors report
The preprint tests spatial reasoning and constrained editing using urban data. Its authors say structured GeoJSON worked better than map images.
Guth briefing · AI-written
Grafana GEM 2.17 adds retries and remote read for cross-cluster queries
The release notes also list a fix for unreported partial results when a remote-read stream breaks.
Guth briefing · AI-written
Upstash expands Workflow labels with match-any filtering
The TypeScript SDK accepts multiple labels per run, while context.labels replaces the deprecated singular field.
Guth briefing · AI-written
Supabase documents a DuckLake destination for Pipelines, still in public alpha
Supabase has published documentation for a DuckLake destination in Supabase Pipelines, its managed service for replicating Postgres tables.
Guth briefing · AI-written
Microsoft outlines 30 Dynamics 365 CRM skills for Copilot and custom agents
The October 8 announcement targets sales and service workflows, while related agents have different preview schedules.
Guth briefing · AI-written
MotherDuck adds an Admin API endpoint to list users and service accounts
The October 8 release note names two filters and paging controls, but leaves permissions and response fields unspecified.
Guth briefing · AI-written
Carter and Thompson reintroduce bill for a CISA AI task force
The proposal would give the agency a year after enactment to assemble an internal group coordinating AI security work and briefing Congress.
Guth briefing · AI-written
Copilot for JetBrains adds default-model controls and an MCP startup switch
Enterprise model defaults preserve explicit user choices. The release also drops support for JetBrains IDE 2025.1.
Guth briefing · AI-written
Sierra announces Poppy protocol draft and 35 more design partners
The proposal would give personal agents customer-approved access to businesses, with one session spanning websites, APIs and company-run agents.
Guth briefing · AI-written
PlanetScale adds branch metrics to its MCP server
The company says developers can retrieve CPU, memory, disk, replication lag and connection figures through an AI assistant without leaving their editor.
Guth briefing · AI-written
Retell announces built-in memory for returning callers
Contact summaries can help agents resume unfinished requests and skip questions already answered. They do not replace live order lookups.
Guth briefing · AI-written
Exa introduces ATLAS to measure what search agents miss
The 547-task benchmark uses an automated answer key to grade completeness, with Exa’s own agent among the systems tested.
Guth briefing · AI-written
Snorkel expands its open AI benchmark commitment to $30 million
The company also plans a team to find flaws in benchmarks and a research fellowship, but has not detailed how or when the commitment will be distributed.
Guth briefing · AI-written
Kiro CLI 2.29.0 adds configurable memory across local V3 chats
Users can disable memory, limit it to reading saved context or allow updates. Corrections and deletions happen through chat.
Guth briefing · AI-written
Andreessen Horowitz announces TypeSafe AI investment
The firm is betting on Jev’s ability to return decisions directly to software, rather than text that code must parse.
Guth briefing · AI-written
KONST says it raised $30 million for AI data-center expansion in Asia
ADATA Technology led the round, according to KONST, whose plans connect infrastructure construction with metered compute access and enterprise usage controls.
Guth briefing · AI-written
Ai2 reports shorter debugging queues with GPU-time budgets
The institute says the 90th-percentile wait fell from two hours to 30 seconds, though a small baseline sample makes the comparison noisier.
Guth briefing · AI-written
Databricks adds request-tag checks and OpenJev judging to Unity Gateway
The two beta additions come with distinct limits: tags do not establish permissions, and OpenJev does not explain its verdicts.
Guth briefing · AI-written
Deno joins Cloudflare and gives Deploy six months before shutdown
The team will end its own runtime development after a year and turn its attention to making Workers easier to run outside Cloudflare.
Guth briefing · AI-written
Claude Managed Agents adds agent-written workflows in beta
Anthropic says its server can execute the generated programs in the background, running agents in phases and combining their results.
Guth briefing · AI-written
OpenProblemBench turns to AI judges for unsolved math and physics
The preprint tests models on 82 unresolved research problems. Its highest reported score is a model-judged solve rate, not a tally of expert-confirmed solutions.
Guth briefing · AI-written
AgentHorizon puts AI judges to the near-miss test
Built from 166 hours of human-recorded computer use, the benchmark asks evaluators to distinguish correct work from a similar task done instead.
Guth briefing · AI-written
Epoch’s automation tests find research judgment lagging behind coding skill
Six models attempted the organization’s work without intervention. The strongest handled well-defined tasks, but a flawed pilot experiment exposed a harder problem.
Guth briefing · AI-written
Buildkite details build comparison and selective retries for its MCP server
The company says its retry tool leaves genuine failures alone and can preview its choices before changing a build.
Guth briefing · AI-written
Anthropic launches free OSS Scanner with unreviewed bug reports
The October 8 announcement also introduced a program offering Claude models, engineers and threat research to critical-infrastructure security providers.
Guth briefing · AI-written
Nullify proposes steering language models away from private answers without retraining
The October 7 preprint reports competitive results on two unlearning benchmarks, but the available abstract does not quantify the advantage.
Guth briefing · AI-written
Gallatin AI announces $50 million Series A to expand military logistics tools
The California company plans engineering hires and product development for software designed to track military readiness and anticipate supply shortages.
Guth briefing · AI-written
Cloudflare lets rules match reported security detection failures
A new field exposes failure reports without changing how detections behave. Existing plan restrictions still apply.
Guth briefing · AI-written
MoE routing logs can sharpen membership attacks, preprint reports
Across nine model-and-data settings, the authors report gains over output-only attacks—even with some forms of restricted telemetry.
Guth briefing · AI-written
Google reports no safety interruptions in 98 supervised AMIE chats
Clinicians reported benefits from the chatbot’s pre-visit summaries, but Google’s brief account leaves the safety thresholds and diagnostic scoring unexplained.
Guth briefing · AI-written
Braintrust brings application logs into its agent-tracing view
The public preview also accepts logs from standalone services. Linking a record to a particular operation requires its trace context.
Guth briefing · AI-written
Multiverse reports accuracy recovery for a smaller, 4-bit GPT-OSS model
The company’s seven-of-nine benchmark result compares its compressed model with a recovered 60B checkpoint—not the original 120B model.
Guth briefing · AI-written
TII introduces Falcon-ASR with a focus on Emirati speech
The five-language transcription model posts promising Arabic results in TII’s evaluations. The institute offers a demo and describes API access and native apps as planned.
Guth briefing · AI-written
DOE plans up to $100 million for Ph.D. fellowship pilot
The program targets about 100 fellows on four-year pathways, with research placements at a national laboratory or user facility and in industry.
Guth briefing · AI-written
Study reports fewer false refusals after paired-prompt training
The political-persuasion experiment also exposed the cost of aggressive refusal training: one setup rejected 74% of safe XSTest prompts.
Guth briefing · AI-written
Reasoning-trace endpoints can match full-trace training, researchers say
A September 2026 paper proposes dropping the middle of training examples, though the available summaries leave the size and scope of the reported gains unclear.
Guth briefing · AI-written
Sierra announces fleming-1 to flag AI voices on phone calls
Businesses can enable the detector for Sierra-built voice agents, the company says. Its announcement provides no accuracy figures or false-positive rates.
Guth briefing · AI-written
Devin adds Claude Haiku 5.5 as a Fusion sidekick
Cognition says Haiku 5.5 beats Sonnet 5 on its proprietary engineering benchmark at about one-eighth of the cost per task.
Guth briefing · AI-written
Netlify adds OpenAI’s beta Decisions API to AI Gateway
The company says Functions can use the OpenAI SDK without managing API keys, with choice answers restricted to developer-defined options.
Guth briefing · AI-written
A preprint proposes measuring AI agents by how efficiently they improve
The authors report divergent learning trajectories despite comparable opportunities—and say the strongest final performer need not be the most efficient learner.
Guth briefing · AI-written
Station agents scored 62.7% in a research rediscovery test, authors report
Researchers withheld results from three ICLR papers, then counted how many findings-based criteria the agents recovered. The abstract leaves key scoring details unexplained.
Guth briefing · AI-written
Copilot Studio adds app building in public preview, with a managed runtime to host the apps
Microsoft has put app creation into public preview in Copilot Studio, so makers can now build apps, workflows and agents in the same product, according to the company's October 7 post on its September 2026 updates.
Guth briefing · AI-written
Cloudflare declares Organizations generally available
The multi-account management layer covers Enterprise customers and MSSP/Distributor partners, while Organization Roles remains in beta.
Guth briefing · AI-written
Google opens SynthID Detector to the public worldwide
Google says the English-language checker covers images, video and audio made with its own AI and partners’ systems. Apple support is still to come.
Guth briefing · AI-written
Microsoft outlines Agent Lightning v1.0 for training agents in their existing harnesses
The researchers report a 14.6-percentage-point gain on SWE-bench Verified in a coding-agent experiment using about 6,000 training samples.
Guth briefing · AI-written
PIR probe aims to tell hidden model knowledge from missing knowledge
A single-author research paper proposes a way to check whether a language model recognizes an answer it does not give.
Guth briefing · AI-written
Retool announces early access to centrally managed workplace agents
Retool says its shared controls check tool requests outside the model, while excluding laptop access, browser automation and desktop control.
Guth briefing · AI-written
GitHub details fine-tuned model for secret detection
The model uses surrounding code to identify likely credentials across secret alerts, push protection and Copilot security reviews.
Guth briefing · AI-written
GitHub adds Claude Haiku 5.5 to Copilot in a gradual rollout
GitHub lists the lightweight Anthropic model across Copilot’s paid plans, with access controls for Business and Enterprise administrators.
Guth briefing · AI-written
Patent-law trial finds better AI-assisted drafts, uneven learning gains
After three months, senior lawyers outperformed controls on a review task without AI. Junior lawyers showed no average advantage.